noble travails: noble liquid dark matter detectors · noble liquids / dark matter rick gaitskell,...
Post on 16-Aug-2020
10 Views
Preview:
TRANSCRIPT
Gaitskell
Noble Travails:
Noble Liquid Dark Matter Detectors
Rick GaitskellParticle Astrophysics Group, Brown University, Department of Physics
(Supported by US DOE HEP)see information at
http://particleastro.brown.edu/http://gaitskell.brown.edu
v1
LUX Dark Matter Collaboration 2007 v01_7mm <>
Dark Matter Theory and Experiment
~1 event kg-1 day-1
~1 event 100 kg-1 yr-1
• SOME SUSY MODELS� [blue] T. Baltz and P. Gondolo,
Markov Chain Monte Carlos. JHEP 0410 (2004) 052, (hep-ph/0407039)
� [red] J. Ellis et al. CMSSM, Phys.Rev. D71 (2005) 095007, (hep-ph/0502001)
� [red crosses] G.F. Guidice and A. Romanino, Nucl.Phys. B699 (2004) 65; Erratum-ibid. B706 (2005) 65, (hep-ph/0406088)
� [blue crosses] A. Pierce, Finely Tuned MSSM, Phys.Rev. D70 (2004) 075006, (hep-ph/0406144)
Current Experimental Sensitivities
Goals in 2-5 years
SUSY Theory
DM Searches - Aspen - 17 February 2006 Rick Gaitskell, Brown University, DOE
Background Challenges
• Search sensitivity (low energy region <<100 keV)� Current Exp Limit < 1 evt/kg/20 days, ~< 10-1 evt/kg/day� Goal < 1 evt/tonne/year, ~< 10-5 evt/kg/day
• Activity of typical Human� ~10 kBq (104 decays per second, 109 decays per day)
• Environmental Gamma Activity in unshielded detector� 107 evt/kg/day (all values integrated 0–100 keV)� This can be easily reduced to ~102 evt/kg/day using 25 cm of Pb
• Moving beyond this� e.g. High Purity Water Shield 4m gives <<1 evt/kg/day� But you have to focus on intrinsic U/Th contamination ppt (10-12 g/g) levels
• Main technique to date focuses on nuclear vs electron recoil discrimination� This is how CDMS II experiment went from 102 -> 10-1 evts/kg/day
• Environmental Neutron Activity� (α,n) from rock 0.1 cm-2 day-1
� Since <8 MeV use standard moderators (e.g. polyethylene, or water, 0.1x flux per 10 cm� Cosmic Ray Muons generate high energy neutrons 50 MeV - 3 GeV which are tough to moderate� Need for depth (DUSEL) - surface muon 1/hand/sec, Homestake 4850 ft 1/hand/month
Techniques for dark matter direct detection
TYPE DISCRIMINATIONTECHNIQUE
TYPICAL EXPERIMENT
ADVANTAGE
IonizationNone
(Ultra Low BG) MAJORANA, GERDASearches for ββ-
decay, dm additional
Solid Scintillatorpulse shape
discriminationLIBRA/DAMA,
NAIADlow threshold, large
mass, but poor discrim
Cryogeniccharge/phononlight/phonon
CDMS, CRESSTEDELWEISS
demonstrated bkg discrim., low
threshold, but smaller mass/higher cost
Liquid noble gaslight pulse shape
discrimination, and/or charge/light
ArDM, LUX, WARP, XENON, XMASS,
XMASS-DM, ZEPLIN
large mass, good bkg discrimination
Bubble chambersuper-heated bubbles/
droplets COUPP, PICASSOlarge mass, good bkg
discrimination
Gas detectorionization track
resolved DRIFTdirectional sensitivity, good discrimination
R.J. Gaitskell, Ann. Rev. Nucl. Par. Sci, 54 (2004) 315
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
Noble Liquids
• Why Noble Liquids?� Nuclear vs Electron Recoil discrimination readily achieved
• Scintillation pulse shapes• Ionization/Scintillation Ratio
� High Scintillation Light Yields• Low energy thresholds can be achieved (although have to pay close attention to how discrimination
behaves with energy)
� Ionization Drift >>1 m, at purities achieved (<< ppm electronegative impurities)� Large Detector Masses are easily constructed and behave well
• Shelf shielding means Inner Fiducial volumes have very low activity (assuming intrinsic activity of target material is low)
— BG models get better the larger the instrument• Position resolution of events very good in TPC operation (ionization)• Dark matter cross section on nucleons goes down at least to σ ~ 10-46 cm2 == 1 event/100 kg/year (in Ge or Xe), so need a large fiducial mass to collect statistics
� Cost & Practicality of Large Instruments• Very competitive / Simply Increase PMTs
• “Dark Matter Sensitivity Scales As The Mass, Problems Scale As The Surface Area”
• Scintillation Light Yield comparable to NaI 40,000 phot/MeV
• liquid rare gas gives both scintillation and ionization signals
Z (A) BP (Tb) at 1 atm [K]
liquid density at Tb [g/cc]
ionization [e-/MeV]
scintillation [photon/MeV]
He 2 (4) 4.2 0.13 39,000 22,000
Ne 10 (20) 27.1 1.21 46,000 30,000
Ar 18 (40) 87.3 1.40 42,000 40,000
Kr 36 (84) 119.8 2.41 49,000 25,000
Xe 54 (131) 165.0 3.06 64,000 46,000
Noble Liquids as detector medium
• liquid rare gas gives both scintillation and ionization signals
• Scintillation is decreased (~factor 2) when E-field applied for extracting ionization
Z (A) BP (Tb) at 1 atm [K]
liquid density at Tb [g/cc]
ionization [e-/keV]
scintillation [photon/keV]
He 2 (4) 4.2 0.13 39 22
Ne 10 (20) 27.1 1.21 46 30
Ar 18 (40) 87.3 1.40 42 40
Kr 36 (84) 119.8 2.41 49 25
Xe 54 (131) 165.0 3.06 64 46
Noble Liquids as detector medium
In LXe ~30% of electron recoil energy appears as scintillation light (7 eV photons)
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
Noble Liquid Comparison (DM Detectors)
Scintillation Light Intrinsic Backgrounds
Ne (A=20)
$60/kg100% even-even nucleus
85 nm Requires wavelength Shifter
Low BP (20K) - all impurities frozen outNo radioactive isotopes
Ar (A=40)
$2/kg(isotope separation >$1000/kg)
~100% even-even
125 nmRequires wavelength shifter
Nat Ar contains ~39Ar 1 Bq/kg == ~150 evts/keVee/kg/day at low energies. Requires isotope separation, low 39Ar source, or very good discrimination (~106 to match CDMS II)
Xe (A=131)$800/kg
50% odd isotope
175 nmUV quartz PMT window
No long lived isotopes. 85Kr can be removed by charcoal or distillation.
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
Noble Liquid Comparison (DM Detectors)
Scintillation Light Intrinsic Backgrounds
Ne (A=20)
$60/kg100% even-even nucleus
85 nm Requires wavelength Shifter
Low BP (20K) - all impurities frozen outNo radioactive isotopes
Ar (A=40)
$2/kg(isotope separation >$1000/kg)
~100% even-even
125 nmRequires wavelength shifter
Nat Ar contains ~39Ar 1 Bq/kg == ~150 evts/keVee/kg/day at low energies. Requires isotope separation, low 39Ar source, or very good discrimination (~106 to match CDMS II)
Xe (A=131)$800/kg
50% odd isotope
175 nmUV quartz PMT window
No long lived isotopes. 85Kr can be removed by charcoal or distillation.
0 20 40 60 80 10010!6
10!5
10!4
10!3
10!2
10!1
100
mWIMP=100 GeV, !W!N =1.0×10!42 cm2
Recoil Energy, Er [keVr]
(das
h)R
ate>
E r [kg/
day]
(lin
e)dN
/dE r [/
keVr
/kg/
day]
Xe A=131Ar A= 40Ne A= 20
Integrated Rates (dash)
Differential Rates (lines)
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
Noble Liquid Comparison (DM Detectors)
Scintillation Light Intrinsic Backgrounds
Ne (A=20)$60/kg
100% even-even nucleus
85 nm Requires wavelength Shifter
Low BP (20K) - all impurities frozen outNo radioactive isotopes
Ar (A=40)$2/kg(isotope separation >$1000/kg)
~100% even-even
125 nmRequires wavelength shifter
Nat Ar contains ~39Ar 1 Bq/kg == ~150 evts/keVee/kg/day at low energies. Requires isotope separation, low 39Ar source, or very good discrimination (~106 to match CDMS II)
Xe (A=131)$800/kg
50% odd isotope
175 nmUV quartz PMT window
No long lived isotopes. 85Kr can be removed by charcoal or distillation.
Scintillation Light Intrinsic Backgrounds WIMP (100 GeV) Sensitivity vs Ge >10 keVr
Ne (A=20)$60/kg
100% even-even nucleus
85 nm Requires wavelength Shifter
Low BP (20K) - all impurities frozen outNo radioactive isotopes
Scalar Coupling: Eth>50 keVr, 0.02x
Axial Coupling: 0 (no odd isotope)
Ar (A=40)$2/kg(isotope separation >$1000/kg)
~100% even-even
125 nmRequires wavelength shifter
Nat Ar contains ~39Ar 1 Bq/kg == ~150 evts/keVee/kg/day at low energies. Requires isotope separation, low 39Ar source, or very good discrimination (~106 to match CDMS II)
Scalar Coupling: Eth>50 keVr, 0.10x
Axial Coupling: 0 (no odd isotope)
Xe (A=131)$800/kg
50% odd isotope
175 nmUV quartz PMT window
No long lived isotopes. 85Kr can be removed by charcoal or distillation.
Scalar Coupling: Eth>10 keVr, 1.30x
Axial Coupling: ~5x (model dep)Xe is 50% odd n isotope 129Xe, 131Xe
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
Noble Liquid Detectors: Mechanism & Experiments
• Single phase - scintillation only� e-ion recombination occurs� singlet/triplet ratio 10:1 nuclear:electron
• Double phase - ionization & scintillation� drift electrons in E-field (kV/cm)
Xe*
+Xe
Xe2*
Triplet27ns
Singlet3ns
2Xe2Xe
175nm175nm
Xe** + Xe
Xe2+
+e-
(recombination)
Xe+
+XeIonisation
Excitation
Electron/nuclear recoil
Single phase(Liquid only)
PSD
Double phase(Liquid + Gas)PSD/Ionization
Xenon ZEPLIN I XMASS
ZEPLIN II+III,XENON,
XMASS-DM, LUX
ArgonDEAP,CLEAN
WARP,ArDM
Neon CLEAN
time constants depend on gase.g. Xe 3/27nsAr 10/1500ns
wavelengthdepends on gase.g. Xe 175nmAr 128nm
Energy Deposition / Partition into various excitations
Nigel Smith, RALThese mechanisms apply to all Nobles
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
CLEAN Ar PSDData taken with Micro-CLEAN (McKinsey, Yale)
Profile of light pulse for electrons and neutrons
SingletTriplet
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
(McKinsey, Yale)
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
miniCLEAN (proposed)
• 100 kg miniCLEAN� WIMP Goal ~5 x10-45 cm2 � 10 events/year
• Backgrounds� PMT Gammas
• Requires better than 10-8 rejection of ER at 50 keVr
• Currently demonstrated 10-5 >50 keVr (limited by neutron bg in lab)
� PMT neutrons• Studies on going, but these are
expected to be limitation to sensitivity of smaller instrument
• Less of problem in larger target � Position Reconstruction
• How well can events leaking from outer
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
DEAP-1 (being deployed)
• DEAP-1 (Boulay / Hime)� Also based on scintillation PSD alone� Queen’s (Boulay) leading effort - Canadian Groups + Yale/LANL� 7 kg LAr with 2x PMT
• Have been studying PSD using tagged 22Na source to limit lab neutron contamination• Preliminary data showing ~10-4-10-5 discrimination. Will continue to push stats.
� Detector will be taken underground at SNOLab shortly� Poor position reconstruction and so likely to be limited by surface events
Mark Boulay Nov 2006
ET 9390 PMT 5”6” acrylic guide
11” x 6” (8” CF) tee
Acrylic vacuum chamber
Quartz windows
poly PMT supports
inner surface 97% diffuse reflector,Covered with TPB wavelength shifter
Neck connects to vacuum andGas/liquid lines
DEAP-1 design
Mark Boulay Nov 2006
DEAP & CLEAN “ULTIMATE” designs
•Design is driven by need for neutron reduction via hydrogenous material•Vacuum thermal insulation versus ice thermal insulation•Ice insulation not the preferred design for neon due to heat loads
•Liquid Argon 87 K (greater than LN2), Liquid Neon (27 K)
DEAP-3“miniCLEAN” 1000 kg
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
XMASS 100 kg (Xe) - Japan
• XMASS� 100 kg Prototype operated
• Limited PMT coverage / Position reconstruction of events near walls at center� Next step is to 800 kg
! Status of 800 kg detector
" Basic performances have been already
confirmed using prototype detector
# Method to reconstruct the vertex and
energy
# Self shielding power
# BG level
" Detector design is going using MC
# Structure and PMT arrangement (812 PMTs)
# Event reconstruction
# BG estimation
" New excavation will be done soon
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
XMASS 800 kg - Japan• 60 triangles
• 10 PMT/triangle x 60 =
600 PMTs
• + 212 PMTs in triangle
boundary region
• Total 812 PMTs
• Photo coverage 67.0%
• Center to photocathode
~45cm
• Fiducial vloume is 25cm
from center.
• PMTs are inside liquid
xenon.
Design of 800kg
1 detector
Summary• XMASS 800kg detector
– 1 ton liquid xenon, 90cm diameter, 60 triangles, 812 PMTs
– BG level 10-4 dru(day-1kg-1kev-1)
– Dark matter search 10-45 cm2
• Detector design by simulation
– Resolution of event reconstruction• 10keV ~3cm 5keV ~5cm at boundary of fiducial
volume
– Background from PMT• 238U, 60Co ~10-5 dru inside fiducial volume
– Water shield for ambient ! and fast neutron• 200cm shield is enough
Background from PMT 238U
• 1.8 x 10-3 Bq/PMT
• <100keV
– 5cm shield ~10-3 dru
– 10cm shield ~10-4 dru
– 20cm shield ~10-5 dru
All volume
5cm self shield 40cm from center10cm shield 35cm20cm shield 25cm
Reconstructed Energy(keV)
10
-51
0-4
10
-31
0-2
10
-1
1
dru(day-1kg-1keV-1)
Decision on funding of 800 kg phase currently be considered
Gaitskell
XENON Event Discrimination: Electron or Nuclear Recoil?
EGC
Cathode
Grid
AnodeEAG
EAG > EGC
Liquid phase
Gas phase
Within the xenon target:
• Neutrons, WIMPs => Slow nuclear recoils => strong columnar recombination
=> Primary Scintillation (S1) preserved, but Ionization (S2) strongly suppressed
• γ, e-, µ, (etc) => Fast electron recoils =>
=> Weaker S1, Stronger S2
PMT Array(not all tubes shown)
Ionization signal from nuclear recoil too small to be directly detected => extract charges from liquid to gas and detect much larger proportional scintillation signal => dual phase
Simultaneously detect (array of UV PMTs) primary (S1) and proportional (S2) light => Distinctly different S2 / S1 ratio for e / n recoils provide basis for event-by-event discrimination.
Challenge: ultra pure liquid and high drift field to preserve small electron signal (~20 electrons) ; efficient extraction into gas; efficient detection of small primary light signal (~ 200 photons) associated with 16 keVr
Light SignalUV ~175 nmphotons
Time
Primary
Proportional
Interaction (WIMP or Electron)
Liq. Surface
e-e-
e-e-
e-e-
e-e-
e-e-
e-e-
Electron Drift~2 mm/µs
0–150 µsdepending on
depth
~40 ns width
~1 µs width
PMT Array(not all tubes shown)
vv
• PSD and secondary scintillation from ionization drift
• WARP (Carlo Rubbia)u 3.2 kg prototype running at
Gran Sassou Preliminary results reportedu 140-kg detector w/800-kg
active veto under construction
• ArDM (Andre Rubbia)u LEMs for ionization readoutu PMTs for primary scintillationu 1 ton prototype in construction
Two-phase Argon Detectors: WARP and ArDM
WARP
100-l detector
WARP Prototype
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
WARP - Dual Methods of Discrimination
• PSD� Nuclear Recoil “Ion” has
larger prompt component as in single phase
• S2/S1� Also have Ionization/
Scintilation
24
Figure 2.
Figure 3.
Figure 4.
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
WARP Recent Results (Jan 07) astro-ph/0701286
• Analysis with no events above 55 keV (energy threshold selected a posteriori) yields limit at cyan line (5x above CDMS).� At this threshold energy Ar is 1/10 as sensitive to WIMPs
per unit mass as Ge E>10 keV� The 40 keVr cyan dashed line is a simple a “what if” there
were no events above 40 keVr
• Have new data run of ~50 kg-days with improved electronics - suggest that it will remove some/all of low energy events. (Announce soon)
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
39Ar Beta Background - Event Rejection vs Removal
• Note that regular Ar contains 39Ar ~1 Bq/kg, which gives beta spectrum (end point ~500 keV) with a low energy tail of ~150 evts/keVee/kg/day
• This means that in order to match current best CDMS II sensitivity an Ar experiment must deliver at least ~106 rejection.� Fiducialization/multiple scatter cuts don’t help in reducing this rate
• Possible ways of dealing with it� Improve discrimination so it become irrelevant (although still have to deal with the
event rate 1 kHz in 1 tonne)� Isotopic reduction (WARP have taken delivery of 3 liters of Ar with ~1/50 activity for
running in WARP prototype)� Extraction of Ar from underground wells
• However, underground (n,p) process in 39K will generate 39Ar. (n > 3 MeV are generated by U/Th decays)
• An initial sample that was tested from an underground well had 50x (larger) than usual 39Ar:Ar concentration - large survey will be required to understand factors effecting levels.
N.J.T.Smith ILIAS/ASPERA Paris Workshop January ‘07
ZEPLIN-II Detector
5 months continuous operation1.0t*day of raw DM data
140 m
m~
75 µ
s
S1
S2
ELXe
GXe
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
Discrimination Power
• AmBe calibration (upper)• Co-60 Calibration (lower)
� Used to define acceptance window� 50% n.r. acceptance shown� lower S2/S1=40 bound fixed� Box defined 5-20keVee
• Uniform population across plots� high rate calibrations (esp Co-60)� coincidences between events and ‘dead-
region’ events
• 98.5% γ discrimination at 50% n.r. acceptance
20
30
405060708090
100
200
300
400500
0 5 10 15 20 25 30 35 40energy, keVee
S2/S
1
20
30
405060708090
100
200
300
400500
0 5 10 15 20 25 30 35 40energy, keVee
S2/S
1
Neutron Calibration
Gamma Calibration
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
ZEPLIN II
• 31 live days running, 225 kg-days exposure� Red Box is 5-20keVee, 50% NR
acceptance based on neutron calibration
� 29 candidate events seen• Estimate 50% from ER leakage from
upper band• Other 50% from lower band which are
RAdon daughters plating on PTFE side walls
� Both populations have been modeled and subtraction performed
� Final results is <10.4 events (90% CL) consistent with WIMP
WIMP Search Data
Additional Population: Rn daughter recoils
Candidates
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
The XENON10 Detector
• 22 kg of liquid xenon� 15 kg active volume� 20 cm diameter, 15 cm drift
• Hamamatsu R8520 1’’×3.5 cm PMTs� bialkali-photocathode Rb-Cs-Sb,� Quartz window; ok at -100ºC and 5 bar� Quantum efficiency > 20% @ 178 nm
• 48 PMTs top, 41 PMTs bottom array� x-y position from PMT hit pattern; σx-y≈ 1 mm� z-position from ∆tdrift (vd,e- ≈ 2mm/µs), σZ≈0.3 mm
• Cooling: Pulse Tube Refrigerator (PTR), • 90W, coupled via cold finger (LN2 for
emergency)
The XENON10 Collaboration
Columbia University Elena Aprile, Karl-Ludwig Giboni, Maria Elena Monzani, Guillaume Plante, Roberto Santorelli and Masaki YamashitaBrown University Richard Gaitskell, Simon Fiorucci, Peter Sorensen and Luiz DeViveirosRWTH Aachen University Laura Baudis, Jesse Angle, Joerg Orboeck, Aaron Manalaysay and Stephan SchulteLawrence Livermore National Laboratory Adam Bernstein, Chris Hagmann, Norm Madden and Celeste WinantCase Western Reserve University Tom Shutt, Peter Brusov, Eric Dahl, John Kwong and Alexander BolozdynyaRice University Uwe Oberlack, Roman Gomez, Christopher Olsen and Peter ShaginYale University Daniel McKinsey, Louis Kastens, Angel Manzur and Kaixuan NiLNGS Francesco Arneodo and Alfredo FerellaCoimbra University Jose Matias Lopes, Luis Coelho, Luis Fernandes and Joaquin Santos
Noble Liquids / Dark Matter Gaitskell / Brown University
XENON10: Ready for Low Background Operation
Installation of the Detector... ...and we are operational
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
XENON10 Live time at Gran Sasso
High Stats Gamma Calibrations
Periodic Gamma Calibs
92% live
BLIND WIMP SEARCH DATA
NOT BLIND WIMP Search Data
Neutron Calibration
WIMP search end
• Discuss data from High Statistics Gamma Calib, Neutron Calib and NON BLIND WIMP search data ~20 live days
• WIMP Search results (from 80 live day) will be announced at April APS Meeting
Noble Liquids / Dark Matter Gaitskell / Brown University
XENON10 Detector
89 PMTs: Hamamatsu R8520-AL 2.5 cm square
R =10cm
z =15cm
Noble Liquids / Dark Matter Gaitskell / Brown University
XENON10 Event Discrimination
Example: Low Energy Compton Scatter • S1=15.4 phe ~ 6 keVee • Drift Time ~38 μs => 76 mm
s1: Primary Scintillation Created by Interaction LXes2: Secondary Scintillation Created by e- extracted & accelerated in GXe
Expect > 99% rejection efficiency of γ/n Recoils…Reduction of Backgrounds =>
Reduction of Leakage Events
Incident Particle
s1e- e- e- e- e-
E = 1kV/cm
s2e- e- e- e-
(s2/s1)ER > (s2/s1)NR
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
Enr =S1Ly
1Leff
Ser
Snr
Xe Recoil Energy [keVr]
10 20 30 40 50 60 70 80 90 100
!L
igh
t Y
ield
, re
lati
ve 1
22 k
eV
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35Akimov 2002
Aprile 2005
Arneodo 2000
Chepel 2005
S1 [keVr]
0 10 20 30 40 50 60 70 80 90 100
Co
un
t ra
te [
dru
]
1
10
210
310
MC
Single nuclear recoils
Xe Recoil Energy [keVr]
10 20 30 40 50 60 70 80 90 100
!L
igh
t Y
ield
, re
lati
ve
12
2 k
eV
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35Akimov 2002
Aprile 2005
Arneodo 2000
Chepel 2005
S1 [keVr]
0 10 20 30 40 50 60 70 80 90 100
Co
un
t ra
te [
dru
]
1
10
210
310
MC
Leff = 0.19
Single nuclear recoils
Nuclear recoil spectrum
Data
MC
Neutron Calibration
Gamma Events in Fid. Region (~5 kg)
Neutron Calibration(Center of distribution)
Response consistent with gaussian at <0.1%
Gamma Calibration (Electron Recoils == Background)
(Best fit would be with QF rising at low energy (see Chepel, QF measurement following slide)
No free params in MC rate & energy scale (fixed based on gamma ref & above ground NR calibs)
See Manzur Talk
Sorensen (Brown)
see K. Ni Talk
Sorensen (Brown)
Sorensen (Brown)
Note: ER and NR curves shown are not final versions used in 58 day WIMP Blind analysis
Fit here assumes QF flat 19%
~20% discrepe
keVr
De Viveiros - Brown University April 2007 v01 <>APS April Meeting
S1
log(
S2
/ S1
)
“Leakage” Events
Applying the Gamma-X Cuts to XENON10 Data
§XENON10 Blind Analysis – 58.6 days§WIMP “Box” defined at
• ~50% acceptance of Nuclear Recoils (blue lines): [Centroid -3σ]
• 2-12keVee (2.2phe/keVee scale)
• Assuming QF 19% 4.5-27 keVr
§ 10 events in the “box” after all primary analysis blind cuts (o)
§ 5 of events are consistent with gaussian tail from ER band
• Fits based on ER calibrations projected 7.0 +2.1-1.0 events
§ 5 of these are not consistent with Gaussian distribution of ER Background
log ( S2 / S1) vs S1“Straightened Y Scale” – ER Band Centroid => 2.5
Removed by PrimaryGamma-X cuts
2 - 12 keVee4.5 - 27 keVr @ QF 19%
See Aprile / Manalaysay Talk
De Viveiros - Brown University April 2007 v01 <>APS April Meeting
S1
log(
S2
/ S1
)
Secondary Gamma-X cuts
Applying the Gamma-X Cuts to XENON10 Data
§XENON10 Blind Analysis – 58.6 days§WIMP “Box” defined at
• ~50% acceptance of Nuclear Recoils (blue lines): [Centroid -3σ]
• 2-12keVee (2.2phe/keVee scale)
• Assuming QF 19% 4.5-27 keVr
§ 10 events in the “box” after all primary analysis blind cuts (o)
§ 5 of events are consistent with gaussian tail from ER band
• Fits based on ER calibrations projected 7.0 +2.1-1.0 events
§ 5 of these are not consistent with Gaussian distribution of ER Background
• 4 out of 5 events removed by Secondary Blind Analysis (looking for missing S2/Gamma-X events)
• Remaining event would have been caught with 1% change in cut acceptance : WIMP SIGNAL UNLIKELY
log ( S2 / S1) vs S1“Straightened Y Scale” – ER Band Centroid => 2.5
“Leakage” Events
Primary Gamma-X cuts
SecondarySignal Quality Cuts
2 - 12 keVee4.5 - 27 keVr @ QF 19%
See de Viveiros Talk
De Viveiros - Brown University April 2007 v01 <>APS April Meeting
§Why are are there fewer events in box in low energy?§Discrimination improves ! at lowest energies - NR and ER bands move apart in log(S2/S1) plot§Missing S2 events less frequent for low energies, (multiple scatters, boost S1)
!"#$%&'()*+,&- .#//#01&23%45+
!"#$%&'()*+&6#4&76689&:79&)8;5<=;%&>6#55%9&%3%45+&?<5@&A&+6#55%9&<4&5@%&>%4+<5<3%&B7;8/% &A&+6#55%9&<4&5@%&C%3%9+%&"<%;D&B7;8/%E&A&+6#55%9&<4&5@%&C%3%9+%&"<%;D&C%F<74! G#9F%&>A&+<F4#;&H&>%4+<5<3%&B7;8/%&>A&I&&C"C&>AJ&
! >/#;;&>K&+<F4#;&H>%4+<5<3%&B7;8/%&>K&74;LJ
! MN&;7?&>KO>A
20:<%;D
%0%0%0>%4+<5<3%
B7;8/%
HAP&6/J! >K&=7+<5<74&D#5#&74;L&:79&+6#55%9+&<4&5@%&+%4+<5<3%&
37;8/%&- 47&<4:79/#5<74&#Q785&+6#55%9+&<4&5@%&C%3%9+%&"<D86<#;&C%F<74
!'%&6#;;&+86@&%3%45+&!.#//#01,E&Q%6#8+%&7:&5@%&84$47?4&67/=74%45&:79&+6#55%9+&<4&
>A
>A
HAP&6/J
C%3%9+%
"<%;D2
0:
6#5@7D%
7:&5@%&84$47?4&67/=74%45&:79&+6#55%9+&<4&C%3%9+%&"<D86<#;&C%F<74
!2;%65974&3+R&S86;%#9&C%67<;&T<+69</<4#5<74&<+&3%9L&D<::<68;5&:79&5@%+%&%3%45+&Q%6#8+%&7:&< ; 5 &>K&< : 5< & 5@ & 5 &
>A
(467/<4F
*#95<6;%
C%F<74
HARK&6/J
:<%;D
<467/=;%5%&>K&<4:79/#5<74&- 5@%+%&%3%45+&/<F@5&Q%&#&+7896%&7:&!;%#$#F%,&H2C&%3%45+&/<F@5&Q%&<D%45<:<%D&#+&SC&%3%45+JR
*#95<6;%
>A >K47&>K&U
T%&B<3%<97+&0 V97?4&W4<3%9+<5L X=9<;&KYYZ&&3YA&[\NX*>&X=9<;&)%%5<4F
Absence of Low Energy Candidate Events (2-7 keVee)
Rejection and Predicted Leakage
A. Manalaysay; April 14, 2007
For each energy bin, the ER
rejection at 50% NR
acceptance is calculated
based on the Gaussian fits tothe bands in !-space.
Shown are the ER rejection
powers for both primary and
secondary analyses.Manalaysay Talk
de Viveiros Talk
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
Setting Limit
See Dahl Talk
CDMS II
XENON10 (sub. bg)
XENON10
• Effect on dm sensitivity associated with varying assumption of “best fit” to nuclear recoil light yield� Low energy QF: assume 19% constant as default� Also consider low energy asymptote >30% - <10%
Yellin Maximum Gap AnalysisPRD 66 (2002) 032005
Allows fit to Sig+Known BG+Unknown BG
Text
Based on 10 events
Based on 10, allowing for 7 BG events
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
XENON10
• In situ Neutron Calibration agreed very closely with calibrations of above ground prototypes.
• High Stats Gamma Calib and Preliminary non-blind WIMP search (20 live days) shows performance very similar performance
• Discrimination (S2/S1) - Behavior very encouraging� Gaussian Component
• Due to Recombination fluctuations, and Poisson stats at lowest energies (<5 keVee)� Non-gaussian (systematics) Contribution
• Non-gaussian “LOW TAIL” component is being eliminated at better than 1000:1— Tail events removed using cuts tuned on gamma calib (but this is NON blind analysis)
• Main Cuts used to — Fiducial Volume - eliminate events at edge Teflon where charge (S2) collection is poor— More than one S2 pulse indicating multiple scatter— S1 light hit pattern unusual - e.g. if most of signal is concentrated in few adjacent
bottom PMTs, indicates additional scattering in Xe below cathode grid
LUX Dark Matter Collaboration 2007 v01_7mm <>
LUX Dark Matter Experiment - Summary• Brown [Gaitskell], Case [Shutt], LBNL [Lesko] , LLNL [Bernstein],
Rochester [Wolfs], Texas A&M [White], UC Davis [Svoboda/Tripathi], UCLA [Wang/Arisaka/Cline]
o XENON10, ZEPLIN II (US) and CDMS; ν Detectors (Kamland/SuperK/SNO/Borexino); HEP/γ-ray astro
o (Also ZEPLIN III Groups after their current program trajectory is established)o Co-spokespersons: Shutt (Case)/Gaitskell (Brown)
• 300 kg Dual Phase liquid Xe TPC with 100 kg fiducialo Using conservative assumptions: >99% ER background rejection for 50% NR
acceptance, E>10 keVr(Case+Columbia/Brown Prototypes + XENON10 + ZEPLIN II)
o 3D-imaging TPC eliminates surface activity, defines fiducial
• Backgrounds:o Internal: strong self-shielding of PMT activity
• Can achieve BG γ+β < 7x10-4 /keVee/kg/day, dominated by PMTs (Hamamatsu R8778 or R8520).
• Neutrons (α,n) & fission subdominant
o External: large water shield with muon veto.• Very effective for cavern γ+n, and HE n from muons
• Very low gamma backgrouns with readily achievable <10-11 g/g purity.
• DM reach: 2x10-45 cm2 in 4 monthso Possible <5x10-46 cm2 reach with recent PMT activity reductions, longer running.
5
Compared to a standard Pb and polyethylene shield, water shield readily achieve much lower gamma backgrounds, and provides superior neutron shielding, especially for the very high energy tail of neutrons from the cavern walls. It is also very cost competitive, particularly as the detector size increases, allows
great flexibility in changing the size and shape of the central detector.
Figure 3.1. (left) Over view of the LUX detector system. Shown are the 6 m Ø water tank with Cherenkov-light muon veto, the central cryostat for liquid Xe, two perpetually cooled Xe storage and recovery vessels, a LN storage vessel, and electronics and equipment racks atop a
work platform. (right) Cutaway view of the LUX detector.
Cryostat and cryogenics
A schematic of the central detector is shown in Fig. 3.1. The cryostat housing the central detector is a simple pair of vacuum insulated cans with the inner housing liquid Xe at 165-190 K and a pressure of up to ~3 bar. The lower portions are made of low background OFHC Cu, whose radioactivity is dominated by cosmogenic activation (primarily 60Co) on the surface; in our background estimates we assume <12 months surface exposure during fabrication. The low background cans are brazed to standard stainless steel (wire-seal, or similar), while all feed-throughs and auxiliary flanges are off-the-shelf, metal-seal, ultra high vacuum components that work at cryogenic temperatures. The stainless steel in this region is some thousand times cleaner than the cavern rock, but also some 100-1000 times hotter than the materials immediately adjacent to the Xe. A simple internal Cu (and possibly polyethylene) shield just underneath
these standard flanges shields their activity.
The cryogenic system faces the need for very high cooling power during cool-down and condensing Xe during gas-phase purification, but very low heat loads otherwise, comparable to a liquid nitrogen dewar. Our coolant is liquid nitrogen, which is both simple, and provides the intrinsic safety of having the system cool in the event of a power loss. We have developed a nitrogen thermosyphon capable of transporting ~1 kW of cooling power using sealed tubing from a detached liquid nitrogen bath to locations in the detector. Using this, we will provide separate cooling power for cool-down and condensing Xe, and will also establish a stable thermal gradient by holding the bottom of the cryostat slightly colder than the top. Static heat loads and additional thermal gradients are minimized by the use of a thermal radiation shield and super-insulation. Active temperature control will be employed on the
http://www.luxdarkmatter.org
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
• The rate of ER events in FV is determined by small angle scattering Compton events, that interact once in the FV� The rate of above events is suppressed by the
tendency for the γ’s to scatter a second time. Either on the way in, or way out.
� The chance of no secondary scatter occurring is more heavily suppressed the more LXe there is
• The important optimization is to maximize the amount of LXe that lies along a line from the greatest sources of radioactivity (PMTs?) that pass through the FV.
• Example for 1.5 MeV γ from outside LXe volume� Energy Spectrum for part of energy deposited in FV� Energy spectrum for all energy in detector� Additional application of multiple scatters cut has little
additional effect on low energy event rate
• Conclusion for Event Suppression� xyz resolution of detector is important simply in defining
FV. Little additional reduction from locating vertices.� (Full xyz hit pattern does assist in bg source
identification)
Topology of Gamma Events That Deposit Energy in FV
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
Scaling LXe Detector: Fiducial BG Reduction /1
• Compare LXe Detectors (factor 2 linear scale up each time)15 kg (ø21 cm x 15 cm) -> 118 kg (ø42 cm x 30 cm) -> 1041 kg (ø84 cm x 60 cm)� Monte Carlos simply assume external activity scales with area (from PMTs and cryostat) using XENON10
values from screening
>102 reduction
x10 reduction
15 kg 118 kg 1041 kgx2 linear x2 linear
Gross Mass
Fiducial Massdru = cts/keVee/kg/day
Low energy rate in FV before any ER vs NR rejection /keVee/kg/day
LUX Dark Matter Collaboration 2007 v01_7mm <>
LUX program: exploit scalability
• LUXcore: Final engineering for large-scale detectoro Cryostat, >100 kV feedthrough, charge drift, light collection over large distance
o Full system integration, including ~1m water shieldo 40 kg narrow “core”, 14 PMTs, 20 cm Ø x 40 cm tall.
• Radial scale-up requires full-funding.o Under construction, Jan 2007, operations at Case: spring 2007.
LUX detector
LUXcore
• LUX in ~ 6m Ø water shield• Very good match to early-implementaion DUSEL (e.g., Homestake “Davis” cavern)
o SNOLAB LOI
• System scalable to very large mass.
LUX Experiment
LUX Dark Matter Collaboration 2007 v01_7mm <>
Water Shield - Homestake - Davis Cavern
Homestake DUSEL
Homestake
Deep Underground Science and
Engineering Laboratory
Noble Liquids / Dark Matter Rick Gaitskell, Brown University, DOE
Homestake / Potential DUSEL Site (Lesko, LBL)
• DUSEL process for new national underground lab.� Site Decision mid 2007 (Full DUSEL lab 2010->)
• 4850 mwe depth at Homestake - early program.• Water Shield: >4 m shielding / 10 module system
14 m 16 m1.75 m
LUX Dark Matter Collaboration 2007 v01_7mm <>
Dark Matter Results and (some of Goals)
• Dark Matter Goalso LUX - Sensitivity curve at 2x10-45 cm2 (100 GeV)
• Exposure: Gross Xe Mass 300 kgLimit set with 120 days running x 100 kg fiducial mass x 50% NR acceptance
—If candidate dm signal is observed, run time can be extended to improve stats
• ~1 background event during exposure assuming most conservative assumptions of ER 7x10-4 /keVee/kg/day and 99% ER rejection
—ER bg assumed is dominated by guaranteed Hamamatsu PMT background (R8778 or R8520) - recent PMTs from Hamamatsu achieving lower backgrounds, but not guaranteed
—Improvements in PMT bg (and rejection power) will extend background free running period, and DM sensitivity
o Comparison• SuperCDMS Goal @ SNOLab: Gross Ge Mass 25 kg
(x 50% fid mass+cut acceptance) Limit set for 1000 days running x 7 SuperTowers
Edelweiss I
ZEPLIN II CDMS IIWARP
LUX (Proj)SuperCDMS@Soudan
SuperCDMS@SNOLab
XENON10
LUX Dark Matter Collaboration 2007 v01_7mm <>
Noble Liquids for Dark Matter• Summary
o Past two years we have seen rapid progress in demonstrated performance (NR-ER discrimination/energy resolution/light yields) of Noble Liquid Detectors in low energy regime
o Competitive WIMP Search Results from WARP (Ar), ZEPLIN II (Xe), XENON10 (Xe)
• Single Phase (Liquid only) - Pulse Shape Discrimination (ER)o Ar/Ne demonstrating >105:1 discrimination at 50 keVr, limitations not fundamental.
• Will push these tests to 108:1 using higher light yields/shielding in test facilities (required for 10-45 cm2 dm reach)
o Position reconstruction based on photoelectron hit patterns (timing not useful in <=10 tonne scale). Mis-reconstruction
• 39Ar (160 evts /keVee/kg/day) / Rn daughters on surfaces (major issue)
• Dual Phase (Liquid Target/Ioniz Readout in Gas) - Discrim. Ionization/Photons+PSD (Ar)o Xe TPC Operation: ZEPLIN II / XENON10 (20-35 kg target)
• Discrimination established ~102:1 (50% NR acceptance), fiducialize to get further bg reduction —Xe intrinsically very low activity (cf XMASS) , so scaling works
o Ar TPC (WARP) - studying use of Ionization + PSD• Discrimination Ionization ~102:1 + PSD >104:1 (energy threshold should be improved with better elec.)
• Scaling of Technologyo Detector WIMP sensitivity improves very significantly with sizeo Designs are very scalable - 1 event/100 kg/month (10-45 cm2) in a few years seems very realizableo Future instruments for 10-46 – 10-47 cm2 also realistic (performance & cost)
top related